Multivoxel 1H-MR Spectroscopy Biometrics for Preoprerative Differentiation Between Brain Tumors
(2018) In Tomography : a journal for imaging research 4(4). p.172-181- Abstract
We investigated multivoxel proton magnetic resonance spectroscopy (1H-MRS) biometrics for preoperative differentiation and prognosis of patients with brain metastases (MET), low-grade glioma (LGG) and high-grade glioma (HGG). In total, 33 patients (HGG, 14; LGG, 9; and 10 MET) were included. 1H-MRS imaging (MRSI) data were assessed and neurochemical profiles for metabolites N-acetyl aspartate (NAA) + NAAG(NAA), Cr + PCr(total creatine, tCr), Glu + Gln(Glx), lactate (Lac), myo-inositol(Ins), GPC + PCho(total choline, tCho), and total lipids, and macromolecule (tMM) signals were estimated. Metabolites were reported as absolute concentrations or ratios to tCho or tCr levels. Voxels of interest in an MRSI matrix were labeled according to... (More)
We investigated multivoxel proton magnetic resonance spectroscopy (1H-MRS) biometrics for preoperative differentiation and prognosis of patients with brain metastases (MET), low-grade glioma (LGG) and high-grade glioma (HGG). In total, 33 patients (HGG, 14; LGG, 9; and 10 MET) were included. 1H-MRS imaging (MRSI) data were assessed and neurochemical profiles for metabolites N-acetyl aspartate (NAA) + NAAG(NAA), Cr + PCr(total creatine, tCr), Glu + Gln(Glx), lactate (Lac), myo-inositol(Ins), GPC + PCho(total choline, tCho), and total lipids, and macromolecule (tMM) signals were estimated. Metabolites were reported as absolute concentrations or ratios to tCho or tCr levels. Voxels of interest in an MRSI matrix were labeled according to tissue. Logistic regression, receiver operating characteristic, and Kaplan-Meier survival analysis was performed. Across HGG, LGG, and MET, average Ins/tCho was shown to be prognostic for overall survival (OS): low values (≤1.29) in affected hemisphere predicting worse OS than high values (>1.29), (log rank < 0.007). Lip/tCho and Ins/tCho combined showed 100% sensitivity and specificity for both HGG/LGG (P < .001) and LGG/MET (P < .001) measured in nonenhancing/contrast-enhancing lesional tissue. Combining tCr/tCho in perilesional edema with tCho/tCr and NAA/tCho from ipsilateral normal- appearing tissue yielded 100% sensitivity and 81.8% specificity (P < .002) for HGG/MET. Best single biomarker: Ins/tCho for HGG/LGG and total lipid/tCho for LGG/MET showed 100% sensitivity and 75% and 100% specificity, respectively. HGG/MET; NAA/tCho showed 75% sensitivity and 84.6% specificity. Multivoxel 1H-MRSI provides prognostic information for OS for HGG/LGG/MET and a multibiometric approach for differentiation may equal or outperform single biometrics.
(Less)
- author
- Durmo, Faris LU ; Rydelius, Anna LU ; Cuellar Baena, Sandra LU ; Askaner, Krister LU ; Lätt, Jimmy LU ; Bengzon, Johan LU ; Englund, Elisabet LU ; Chenevert, Thomas L ; Björkman-Burtscher, Isabella M LU and Sundgren, Pia C LU
- organization
-
- BioCARE: Biomarkers in Cancer Medicine improving Health Care, Education and Innovation
- Diagnostic Radiology, (Lund)
- Neuroradiology (research group)
- Brain Repair and Imaging in Neural Systems (BRAINS) (research group)
- Radiology Diagnostics, Malmö (research group)
- StemTherapy: National Initiative on Stem Cells for Regenerative Therapy
- Neurosurgery
- Tumor microenvironment
- MultiPark: Multidisciplinary research focused on Parkinson´s disease
- Lund University Bioimaging Center
- publishing date
- 2018-12
- type
- Contribution to journal
- publication status
- published
- subject
- in
- Tomography : a journal for imaging research
- volume
- 4
- issue
- 4
- pages
- 10 pages
- publisher
- Grapho Publications LLC
- external identifiers
-
- pmid:30588503
- scopus:85076133525
- ISSN
- 2379-1381
- DOI
- 10.18383/j.tom.2018.00051
- language
- English
- LU publication?
- yes
- id
- 24bbecbe-2e76-42f5-9138-f20255ce3f65
- date added to LUP
- 2019-02-14 09:58:34
- date last changed
- 2024-10-01 16:55:22
@article{24bbecbe-2e76-42f5-9138-f20255ce3f65, abstract = {{<p>We investigated multivoxel proton magnetic resonance spectroscopy (1H-MRS) biometrics for preoperative differentiation and prognosis of patients with brain metastases (MET), low-grade glioma (LGG) and high-grade glioma (HGG). In total, 33 patients (HGG, 14; LGG, 9; and 10 MET) were included. 1H-MRS imaging (MRSI) data were assessed and neurochemical profiles for metabolites N-acetyl aspartate (NAA) + NAAG(NAA), Cr + PCr(total creatine, tCr), Glu + Gln(Glx), lactate (Lac), myo-inositol(Ins), GPC + PCho(total choline, tCho), and total lipids, and macromolecule (tMM) signals were estimated. Metabolites were reported as absolute concentrations or ratios to tCho or tCr levels. Voxels of interest in an MRSI matrix were labeled according to tissue. Logistic regression, receiver operating characteristic, and Kaplan-Meier survival analysis was performed. Across HGG, LGG, and MET, average Ins/tCho was shown to be prognostic for overall survival (OS): low values (≤1.29) in affected hemisphere predicting worse OS than high values (>1.29), (log rank < 0.007). Lip/tCho and Ins/tCho combined showed 100% sensitivity and specificity for both HGG/LGG (P < .001) and LGG/MET (P < .001) measured in nonenhancing/contrast-enhancing lesional tissue. Combining tCr/tCho in perilesional edema with tCho/tCr and NAA/tCho from ipsilateral normal- appearing tissue yielded 100% sensitivity and 81.8% specificity (P < .002) for HGG/MET. Best single biomarker: Ins/tCho for HGG/LGG and total lipid/tCho for LGG/MET showed 100% sensitivity and 75% and 100% specificity, respectively. HGG/MET; NAA/tCho showed 75% sensitivity and 84.6% specificity. Multivoxel 1H-MRSI provides prognostic information for OS for HGG/LGG/MET and a multibiometric approach for differentiation may equal or outperform single biometrics.</p>}}, author = {{Durmo, Faris and Rydelius, Anna and Cuellar Baena, Sandra and Askaner, Krister and Lätt, Jimmy and Bengzon, Johan and Englund, Elisabet and Chenevert, Thomas L and Björkman-Burtscher, Isabella M and Sundgren, Pia C}}, issn = {{2379-1381}}, language = {{eng}}, number = {{4}}, pages = {{172--181}}, publisher = {{Grapho Publications LLC}}, series = {{Tomography : a journal for imaging research}}, title = {{Multivoxel 1H-MR Spectroscopy Biometrics for Preoprerative Differentiation Between Brain Tumors}}, url = {{http://dx.doi.org/10.18383/j.tom.2018.00051}}, doi = {{10.18383/j.tom.2018.00051}}, volume = {{4}}, year = {{2018}}, }